Research map / 3D vision / Human pose estimation
FlashCap: Millisecond-Accurate Human Motion Capture via Flashing LEDs and Event-Based Vision
Zekai Wu, Shuqi Fan, Mengyin Liu, Yuhua Luo, Xincheng Lin, Ming Yan, Junhao Wu, Xiuhong Lin, Yuexin Ma, Chenglu Wen, Lan Xu, Siqi Shen and 1 more
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| use events as residuals on RGB anchors (ResPose) instead of ViTPose with linear interpolation | pose estimation · FlashMotion HPE | MPJPE: 10.06 -> 5.66 | Improved |
| replace spline with linear interpolation of ViTPose (ablation) | ViTPose · FlashMotion HPE | MPJPE: 10.13 -> 10.06 | No clear effect |
| use SNN-CNN hybrid encoder instead of ANN variant (ablation) | ResPose · FlashMotion HPE | MPJPE: 8.12 -> 5.66 | Improved |
| use events plus residual RGB anchoring vs RGB-only ViTPose (ablation) | timing estimation · FlashMotion PMT (Kicking) | Mean timing error (ms): 48.5 -> 7.2 | Improved |
| use events plus residual RGB anchoring vs RGB-only ViTPose (ablation) | timing estimation · FlashMotion PMT (Punching) | Mean timing error (ms): 62.3 -> 4.8 | Improved |
| use pure event-based method (Hybrid ANN-SNN) vs RGB ViTPose | timing estimation · FlashMotion PMT (Punching) | Mean timing error (ms): 62.3 -> 54.1 | Mixed |
| add on-off time distance d_t to LED-cluster matching (ablation) | annotation pipeline · Kicking annotation | Precision: 43.34% -> 99.99% | Improved |
| add period distance d_p to LED-cluster matching (ablation) | annotation pipeline · Kicking annotation | Precision: 69.70% -> 99.99% | Improved |
| add outlier filtering (ablation) | annotation pipeline · Crossing Hands annotation | Precision: 83.17% -> 99.99% | Improved |
| add tracking (ablation) | annotation pipeline · Crossing Hands annotation | Precision: 84.29% -> 99.99% | Improved |
| add outlier filtering (ablation) | annotation pipeline · Kicking annotation | Recall: 95.69% -> 98.99% | Improved |
Closest papers by meaning
- EventCap: Monocular 3D Capture of High-Speed Human Motions Using an Event Camera (CVPR 2020 · similarity 0.89)
- RELI11D: A Comprehensive Multimodal Human Motion Dataset and Method (CVPR 2024 · similarity 0.87)
- HybridCap: Inertia-Aid Monocular Capture of Challenging Human Motions (AAAI 2023 · similarity 0.86)
- MotionPRO: Exploring the Role of Pressure in Human MoCap and Beyond (CVPR 2025 · similarity 0.86)
- Optimal-state Dynamics Estimation for Physics-based Human Motion Capture from Videos (NeurIPS 2024 · similarity 0.84)
- MMVP: A Multimodal MoCap Dataset with Vision and Pressure Sensors (CVPR 2024 · similarity 0.84)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
